Philadelphia Reflections

HSA Lifetime Payments: Limits of Feasibility

The two traditional ways to pay for healthcare are paying directly with cash, or paying indirectly with insurance. Each has advantages, and we return to them later. This book proposes a third payment method which ought to be cheaper, while medical care itself ought to be unaffected. The payment idea grows out of a quirk of modern health care: Children up to age 20 consume only 8% of current medical costs; 92% of healthcare expenses arise decades later. Most families of young people could save up money during the long low-expense period, adding extra compound interest to use later. If this approach reduces overall costs, some of the saving could be used to subsidize the poor. No one doubts some extra interest could be gathered. The really critical question then sharpens: Is it enough to be worth the trouble?

The calculation is not an easy one. In the past century, the nature and cost of healthcare changed dramatically and will change more in the future. Nevertheless, attempts are often made to estimate national health costs; lifetime costs are now widely accepted to range around $350,000 per person, in the year 2000 dollars. Women cost about $50,000 more than men. That's partly a result of the statistical convention of attributing all costs of pregnancy to the mother, and it also reflects females living longer than males. These are daunting amounts of money, but at least we can estimate some upper limit to costs from them. At the other boundary, we know some interest could be earned on almost any balance. So the problem has a solution. The real feasibility issue is whether it produces enough savings to be worth the trouble.

The ability to save varies considerably between families, interest rates vary, longevity increases; no one can know precisely what health care will cost when newborns of today live to be a hundred. On the other hand, estimating national totals is often easier. Dividing national data by the size of the population generates individual averages, which are more natural to comprehend. Furthermore, sometimes we know the available revenue but not the costs. It seems a little cynical to say so, but since one limits the other, they are often (roughly) interchangeable. The rest is a little speculative, and sometimes you just have to make an educated guess.

What a hypothetical average person could afford to pay is one of those speculative matters, and what the average person would willingly pay is even harder to guess. But there are limits to reasonableness; some boundaries can be recognized. So, let's now test what the plausible limits might be, starting with a range of interest rates. As a further preliminary, present longevity is to age 83, and one plausible guess about where it might go next century is age 93. The limit to what almost everyone could afford for a newborn child is guessed to be $500; it seems to be a bargain the government would readily accept as a subsidy to the poor, in order to cover a $350,000 expense. Table # 1 displays the first stab at an estimate. It leads to a conclusion: the proposal of pre-funding health care seems feasible enough under certain circumstances, to justify further investigation.

Interest Return According to Roger Ibbotson, the acknowledged authority on investment statistics, inflation has closely approximated 3% per year for the past century. United States stock market assets have appreciated in a range from 10% (large-cap stocks) to 12.7% (for small-caps) for a century. Growth stocks and value stocks have followed different cycles, but over a span of a century have generated almost identical returns. This table makes the hypothetical assumption that average parents already contribute as much as they can at the birth of their child, and all further additions to the child's fund our investment income. Can the cost of a lifetime of health care be supported by a $500 contribution at birth? Under certain imagined circumstances, the answer seems to be a tentative "Yes" -- if the fund can be invested at 7 to 8%, or the average longevity is between 83 and 93 years. Although it may take a little explanation, these do not seem like unreasonable expectations. It might, therefore, be said that if there are no interruptions or withdrawals (a totally unreasonable expectation by the way), the presently expected cost of an average lifetime of healthcare could be accumulated from the investment of $500 at birth. With no further expenditures than the original $500, although it may be a little too early in the discussion for a skeptical reader to accept that. How about this for an alternative: Although the devil is most assuredly in the details, the goal of paying for a substantial amount of healthcare in this way, is at least conceivable.

Having said that, it should also be firmly stated that paying for all of healthcare costs this way, is neither necessary nor probably even desirable. In the first place, when you make things totally free, they lose their perceived value in the eyes of the recipient. He treats the gift as worthless and is induced to spend money even more carelessly than he does with insurance. Secondly, by placing a cap on the upper bound, we adopt indemnity principles of shifting the risk to the person in control of them. It thus removes the temptation to favor inflation as a way of escaping from debts. Third, the explanation acquires specific numbers to replace vague promises. So, let's set the far more realistic goal of paying for half of it, and seeing if that seems even more feasible by using somewhat reduced limits.

The achievement of $175,000 cannot be made by simply cutting one of the ingredients in half, because that reduces his balance(and its resultant later income) by half, also. The result is the recipient soon gets into a downward spiral, just as the miraculously enhanced income sent his spiraling balance upward. We develop a family of curves (figures 2a to 2d) for different contribution levels in the next chapter, but must first digress to meet an unexpected development There's a sweet spot, and we are already close to it. To test this point, we have some very rough estimates from the AHRQ (the Agency for Healthcare Research and Quality) of the distribution of average health costs by age. If we subtract these costs from the data already mentioned, we would choose 8% as the most likely income, and age 88 as the most likely average longevity. The results are seen in Graph #1 and summarized in Table #2. To assist in the verification, the AHRQ data show that 8% of costs are in the age group 0-20; 13% in age 21-39; 31% in the age group 40-64; and 49% are over 65. The preliminary results are seen in Figure 1a.

Whoops! It is immediately obvious our preliminary description has forgotten something important. We will correct the graph in a minute, but first, we must explain something about paying for healthcare. All of the revenue for healthcare must be generated during the working years of, roughly, 18 to 66 years of age. Ignoring a few trust-fund exceptions, the costs of childbirth, neonatal costs, and childhood are currently borne by the parents of a child. To some extent, the fall-back costs of the grandparents are also covered by people in the working age group. To go even further, when the government pays for dependent healthcare, this too is covered by taxes, which are generated by working people. To summarize, no matter what the direct source, ultimately all healthcare costs are derived from working-age earnings.

In table 1b, we remove the 8% of health costs generated by children, as well as the $44 in revenue which is their portion of the $500 seed money, and redraw the graph; we move the beginning of the revenue curve to the time of birth, gaining three doublings of revenue. The anomalous excess of costs over revenues is reduced but not eliminated. The expected surplus appears as promised, but at the end of life, where it becomes considerably enhanced. The general financial idea is vindicated, but what of the children? Their costs are incorporated into an independent Health Savings Account. The legitimacy of doing so, and the financial consequences, are discussed in the following section. For the present, we can see from Figure 3a. that this approach helps but does not entirely reconcile the financing.

The revenue for that account is introduced at age 35 when health costs are predictably low, and ownership is transferred from the parents to the child. That is, we recognize the validity of such a transfer of responsibility, but during the transition from one system to the next, must yield to the requirements of transition. Because of the overlaps, it may well be desirable to keep the two funds separate for quite a long time. It would seem premature to anticipate dynamic effects on the culture of marriage, divorce, and multiple family health insurance plans, and let the consolidation of accounts remain optional until much later in the unfolding of this scheme. Figure 1b illustrates the effect of consolidating accounts in figure 5c, and considerable experience with this issue probably exists within the life insurance industry. A reverse case can even be made for splitting the account into smaller age subgroups by logical age groups, as a way of easing the entanglements which people get themselves into.

Originally published: Thursday, February 12, 2015; most-recently modified: Tuesday, May 21, 2019